Views
5 months ago

survey-mac-layer

5 Compared to TDMA and

5 Compared to TDMA and CDMA, FDMA is less suitable for operation with low cost devices. The first reason for this is FDMA capable nodes require additional circuitry to communicate and switch between different radio channels. The complicated band pass filters required for this operation are relatively expensive. Another disadvantage of FDMA that limits its practical use is the rather strict linearity requirement on the medium. C. Hybrid MAC Protocols Contention-based protocols adapt easily to changing network scenarios and are better suited for networks with low loads. On the other hand, contention-free protocols eliminate collisions and have better channel utilization at higher loads. To harness the advantages of both classes of protocols, hybrid protocols have been proposed that combine aspects of contention-based and contention-free protocols. 1) Hybrid TDMA/FDMA/CDMA Protocols with Contention: Hybrid MAC protocols proposed in literature usually combine elements of CSMA with TDMA, FDMA and CDMA. Protocols that combine TDMA and CSMA such as [33], [34] behave as CSMA at low contention levels and switch to TDMA type operation at high contention levels. Protocols such as the hybrid MAC (HyMAC) protocol proposed in [35] combine CSMA with TDMA and FDMA where nodes are assigned a frequency as well as a time slot to transmit data once they successfully send a bandwidth request using contention based transmission. Similar protocols where CSMA based bandwidth requests are used to determine the allocation of slots and codes have also been proposed [36], [37], [38]. 2) M2M Communications and Hybrid Protocols: Hybrid protocols address some of the performance issues that arise with contention-based and contention-free protocols. Protocols that switch between random access based operation at low loads and scheduled access at high loads avoid the degraded throughput and collisions of random access protocols at high loads and low channel utilization of scheduled access at low loads. Consequently, hybrid protocols are a promising approach for designing MAC protocols for M2M communications. The main drawback of hybrid protocols that have been proposed in the context of wireless ad hoc and sensor networks is their scalability. Many scenarios with M2M communications have node densities that are an order of magnitude (and more) greater than currently deployed wireless networks. At such high densities, the incidence of collisions during the random access based slot/code/frequency reservation stage of hybrid protocols becomes the bottleneck that prevents the network from achieving a high utilization. Another limiting factor of hybrid protocols for ad hoc and sensor networks when applied to M2M networks is the overhead associated with reconfiguring the system settings in order to accommodate varying traffic conditions and number of active nodes. This limitation is most prominent in the case of TDMA based protocols where the frame length needs to be dynamically adjusted. Such dynamic tuning of frame lengths usually results in the waste of some slots and also requires Centrallized Contention Based ATL-SMACA [43] CERA [44] FASA [46] M2M LTE [47],[48],[52] SIC-ALOHA[54] FS-ALOHA & CTA [56] Distributed IEEE 802.16 M2M [53] Centrallized M2M MAC Protocols Hybrid DPCF-M[39] CSMA-TDMA Hybrid[40][41] IEEE 802.11ah [45] Cognitive [49] Reservation FS-ALOHA[55] Fig. 1. Taxonomy of M2M MAC protocols. Distributed Multichannel [46] Centrallized Contention Free Distributed Cognitive Polling [51] control overhead, in addition to the need for synchronization among nodes. Hybrid protocols based on FDMA and CDMA have better scalability than pure FDMA and CDMA. However, the drawback of FDMA in terms of the requirement for expensive hardware and that of CDMA for complex operation and the need for power control still remains. Consequently, TDMA based hybrid protocols are the most promising of the hybrid protocols in the context of M2M communications. IV. MAC PROTOCOLS SPECIFIC TO M2M COMMUNICATIONS In order to address the unique requirements of M2M communications, an intuitive approach is to develop MAC protocols specific to these environments. Recent research along these lines has proposed various protocols and this section reviews these protocols. A taxonomy of the protocols surveyed in this section is shown in Figure 1 and their comparison in terms of the M2M communication requirements listed in Section II is given in Table I. A. DPCF-M A hybrid MAC protocol for M2M communications named DPCF-M (Distributed Point Coordination Function-M) was proposed in [39] to address energy constrained M2M communication. This protocol uses a hybrid of CSMA/CA and Point Coordination Function (PCF) of IEEE 802.11 for channel access. DPCF-M is designed for scenarios where there are two types of devices: local M2M nodes and gateway-capable nodes. Gateway-capable nodes are equipped with a short range interface for local communication and a cellular radio interface, whereas M2M nodes are equipped with only a low-power short-range radio. For local communication among neighboring nodes, the protocol uses CSMA/CA non beacon mode of the IEEE 802.15.4. However when a M2M node needs to contact an external server through the cellular network, it uses one of the gateway-capable nodes to send the data. The operation of DPCF-M protocol is shown in Fig. 2. In this figure, device 1 is a M2M node and wishes to send

6 Device 1 Device 2 (Dual RAT) Device N RFG IFS B DATA DL SubSlot ACK Slot 1 DATA Device 2 becomes master and the rest become slaves All the devices are in idle mode ACK sleep sleep sleep Sleep DATA UL SubSlot DPCF−M Frame Fig. 2. DPCF-M protocol frame structure [39]. Frame 1 Frame 2 COP NP AP D 1 D 2 ... D M TOP ... ... Slot N ACK UL SubSlot B Time+ All the devices revert to idle mode Frame K AP: Announcement Period NP: Notification Period COP: Contention Only Period TOP: Transmission Only Period Fig. 3. Frame structure for the contention-TDMA hybrid MAC protocol from [40]. data to an external server through the cellular link. Device 1 first obtains access to the local channel using CSMA/CA, and then sends a Request for Gateway (RFG) packet to its selected gateway (device 2, which is a gateway-capable node). On receiving the RFG packet, device 2 assumes the role of a master and starts a temporary cluster by periodically transmitting a beacon during the existence of the cluster. Devices that overhear this beacon (say devices 1, 3 · · · , N) enter into the slave mode. Devices in slave mode suspend their CSMA/CA based operation and transmit only when permitted by the master. The time between two successive beacons is divided into a number of slots and each slot is further divided into uplink and downlink sub-slots. Devices in the cluster are assigned individual slots by the master thereby allowing them to sleep at other times, and nodes without data to send stay silent in their slots. In Fig. 2, it is assumed that both device 1 and device N have data to transmit to the gateway and they are assigned slots 1 and N, respectively. The DPCF-M protocol outperforms CSMA/CA based protocols in terms of the throughput and energy efficiency. However, the energy savings come with additional hardware costs for gateway nodes that require two radios. Also, the protocol does not eliminate the collisions that result during local communication using CSMA/CA. B. Scalable Hybrid MAC A CSMA-TDMA hybrid MAC protocol for M2M communications is proposed in [40]. The protocol divides time in frames and each frame consists of four parts: notification period (NP), contention only period (COP), announcement period (AP), and transmission only period (TOP), as shown in Fig. 3. Each frame starts with a NP where the base station (BS) announces the start of the COP to all nodes. During the COP, nodes with data use p-persistent CSMA to send transmission requests to the BS. Successful nodes are allocated slots to transmit data in the TOP and the nodes are informed of their slots during the AP. The length of the COP may vary from frame to frame. An optimization problem is solved by the BS to determine the optimum COP length and the number of devices that are allowed to transmit in the TOP. The length of the contention period as well as the optimum contention probability for the p-persistent CSMA is communicated to all nodes by the BS during the NP. The protocol in [40] is extended in [41] with the addition of quality of service provisioning and fairness by allowing nodes to choose their contention probabilities according to their priority and observed throughput. While the protocols in [40], [41] incur additional delays and energy consumption due to the time required for the COP and the need for contention, they provide a tradeoff between the performance of p-persistent CSMA and TDMA. C. Adaptive Multichannel Protocol for Large-Scale M2M A contention-FDMA hybrid MAC protocol based on the use of a common control channel is proposed in [42] for use in large scale M2M networks. In the proposed protocol, the available bandwidth is split into a number of channels, with one of them used as the control channel. Also, time is divided into intervals of fixed length and each interval is further divided into three phases: estimation, negotiation and data transmission. The estimation phase consists of a number of time slots in which nodes transmit busy tones on the common control channel if they have data to send or if they hear a busy tone from other nodes, with decreasing probability in each time slot. Based on the total number of busy tones sent and heard, a methodology for statistically estimating the number of active nodes is presented. The negotiation phase consists of a number of slots, and nodes transmit data transmission requests (in the control channel) in each slot with a given probability. Nodes that successfully receive request messages reply back confirming the channel to be used for the data transfer. The length of the negotiation phase as well as the access probability are determined from the estimated number of active nodes. Nodes that have successfully reserved a channel with their receiver in the negotiation phase proceed to transmit their data in the data transmission phase. While the protocols performs well in terms of channel utilization, it adds an extra overhead due to the estimation phase. Also, if all nodes are not in the range of each other, estimates of the number of active nodes and thus the parameters for the negotiation phase may not be the same at all nodes. D. Adaptive Traffic Load Slotted MACA For M2M networks with nodes incapable of carrier sensing, an extension of the slotted MACA protocol called Adaptive Traffic Load slotted MACA (ATL S-MACA) protocol has been proposed in [43]. The ATL S-MACA protocol slightly modifies the basic RTS-CTS-DATA-ACK based scheme of MACA and RTS contention is adaptively controlled based on an estimate of the traffic load. The basic idea behind ATL S-MACA is the observation that slotted MACA reaches its maximum throughput at some value of traffic load G opt

Design and Implementation of MAC Layer of WiFiRe protocol
SURVEY
survey
Surveying
Surveying
SURVEY
SURVEY
Surveying
Survey
SURVEY
Survey
survey